zone has not historically filtered the Internet, while the predominant ISP for the rest of the
country has consistently filtered the Internet.
Modifications can be made to the blocking efforts of a country by the authorities at any
time. Sites can be added or removed at their discretion. For example, during our tests in Iran
the Web site of the New York Times was blocked, but for only one day. Some countries
have also been suspected of introducing temporary filtering around key time periods such as
elections.
Hosting modifications can also be made to a blocked site resulting in it becoming accessible or inaccessible. For example, while Blogspot blogs were blocked in Pakistan due to IP
blocking, the interface to update one’s blog was still accessible. However, Blogspot has since
upgraded its service and the new interface is hosted on the blocked IP, making the interface
to update one’s blog inaccessible in Pakistan. The reverse is also possible. For example, if the
IP address of a Web site is blocked, the Web site may change its hosting arrangement in
order to receive a new IP address, leaving it unblocked until the new IP address is discovered
and blocked.
Summary Measures of Internet Filtering
To summarize the results of our work, we have assigned a score to each of the countries we
studied. This score is designed to reflect the degree of filtering in each of the four major thematic areas: 1) the filtering of political content, 2) social content, 3) conflict- and securityrelated content, and 4) Internet tools and applications. Each country is given a score on a
four-point scale that captures both the breadth and depth of filtering for content of each thematic type (see table 1.5).
Pervasive filtering is defined as blocking that spans a number of categories while blocking
access to a large portion of related content.
Substantial filtering is assigned where either a number of categories are subject to a medium level of filtering in at least a few categories or a low level of filtering is carried out
across many categories.
Selective filtering is either narrowly defined filtering that blocks a small number of specific
sites across a few categories, or filtering that targets a single category or issue.
Suspected filtering is assigned where there is information that suggests that filtering is
occurring, but we are unable to conclusively confirm that inaccessible Web sites are the
result of deliberate tampering.
The scores in table 1.5 are subjective evaluations based upon the quantitative information
gathered during a year of testing and research. In 2006, we tested thousands of Web sites
across more than 120 ISPs in 40 countries, creating a database with close to 200,000 observations. Each observation is in turn based on the conclusion of an average of ten accessibility tests. Despite the breadth of this data, a purely quantitative reporting might be
18
Robert Faris and Nart Villeneuve
country has consistently filtered the Internet.
Modifications can be made to the blocking efforts of a country by the authorities at any
time. Sites can be added or removed at their discretion. For example, during our tests in Iran
the Web site of the New York Times was blocked, but for only one day. Some countries
have also been suspected of introducing temporary filtering around key time periods such as
elections.
Hosting modifications can also be made to a blocked site resulting in it becoming accessible or inaccessible. For example, while Blogspot blogs were blocked in Pakistan due to IP
blocking, the interface to update one’s blog was still accessible. However, Blogspot has since
upgraded its service and the new interface is hosted on the blocked IP, making the interface
to update one’s blog inaccessible in Pakistan. The reverse is also possible. For example, if the
IP address of a Web site is blocked, the Web site may change its hosting arrangement in
order to receive a new IP address, leaving it unblocked until the new IP address is discovered
and blocked.
Summary Measures of Internet Filtering
To summarize the results of our work, we have assigned a score to each of the countries we
studied. This score is designed to reflect the degree of filtering in each of the four major thematic areas: 1) the filtering of political content, 2) social content, 3) conflict- and securityrelated content, and 4) Internet tools and applications. Each country is given a score on a
four-point scale that captures both the breadth and depth of filtering for content of each thematic type (see table 1.5).
Pervasive filtering is defined as blocking that spans a number of categories while blocking
access to a large portion of related content.
Substantial filtering is assigned where either a number of categories are subject to a medium level of filtering in at least a few categories or a low level of filtering is carried out
across many categories.
Selective filtering is either narrowly defined filtering that blocks a small number of specific
sites across a few categories, or filtering that targets a single category or issue.
Suspected filtering is assigned where there is information that suggests that filtering is
occurring, but we are unable to conclusively confirm that inaccessible Web sites are the
result of deliberate tampering.
The scores in table 1.5 are subjective evaluations based upon the quantitative information
gathered during a year of testing and research. In 2006, we tested thousands of Web sites
across more than 120 ISPs in 40 countries, creating a database with close to 200,000 observations. Each observation is in turn based on the conclusion of an average of ten accessibility tests. Despite the breadth of this data, a purely quantitative reporting might be
18
Robert Faris and Nart Villeneuve